In [ ]:
## stats on training data
plt.plot(history.history['loss'], label = 'training loss')
plt.plot(history.history['accuracy'], label = 'training accuracy')
plt.grid(True)
plt.legend()
Out[ ]:
<matplotlib.legend.Legend at 0x203c6226710>
import os
import cv2
os.chdir("D:/Dataset/cancer")
img =cv2.imread('D:/Dataset/cancer/1.jpg')
import os
import cv2
os.chdir("D:/Dataset/Non-Cancer")
img =cv2.imread('D:/Dataset/Non-Cancer/1.jpg')
cv2.imshow('img',img)
cv2.waitKey(6000)
cv2.destroyAllWindows()
resize=cv2.resize(img,(256,256),interpolation=cv2.INTER_AREA)
cv2.imshow('res',resize)
cv2.waitKey(2000)
cv2.destroyAllWindows()
cv2.imwrite("img.jpg",resize)
True
for j in range(1,8):
for i in range(0,50):
im=cv2.imread("D:/Dataset/cancer/"+str(j)+".jpg")
resize=cv2.resize(im,(1028,1028),interpolation=cv2.INTER_AREA)
cv2.imwrite("D:/Dataset/cancer/"+str(j)+str(i)+".jpg",resize)
for j in range(1,6):
for i in range(0,50):
im=cv2.imread('D:/Dataset/Non-Cancer/'+str(j)+'.jpg')
resize=cv2.resize(im,(1028,1028),interpolation=cv2.INTER_AREA)
cv2.imwrite("D:/Dataset/Non-Cancer/"+str(j)+str(i)+".jpg",resize)
for i in range(1,8):
n=str(i)
img=cv2.imread('D:/dataset/cancer/'+n+'.jpg')
resize=cv2.resize(img,(1028,1028),interpolation=cv2.INTER_AREA)
cv2.imwrite("D:/Dataset/Resized cancer/"+n+".jpg",resize)
for i in range(1,6):
n=str(i)
img=cv2.imread('D:/dataset/Non-Cancer/'+n+'.jpg')
resize=cv2.resize(img,(1028,1028),interpolation=cv2.INTER_AREA)
cv2.imwrite("D:/Dataset/Resized Non-Cancer/"+n+".jpg",resize)
from tensorflow.keras.utils import array_to_img, img_to_array, load_img
from keras.preprocessing.image import ImageDataGenerator
datagen = ImageDataGenerator(rotation_range=40,width_shift_range=0.1,
height_shift_range=0.1,shear_range=0.2,
zoom_range=0.2,horizontal_flip=False,
fill_mode='nearest')
def img_gene(img):
img = load_img(img)
x = img_to_array(img)
x = x.reshape((1,) + x.shape)
i = 0
for batch in datagen.flow(x, batch_size=1,save_to_dir='D:/Dataset/cancer',
save_prefix='10', save_format='jpg'):
i+= 1
if i > 20:
break
for p in range(1,7):
num=str(p)
img="D:/Dataset/Resized cancer/"+num+".jpg"
img_gene(img)
import os
# assign directory
directory = "D:/Dataset/Resized Non-Cancer/"
# iterate over files in
# that directory
for c in os.listdir(directory):
f = os.path.join(directory, c)
if os.path.isfile(f):
#img="D:/Dataset/Resized cancer/"+num+".jpg"
img_gene(f)
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import tensorflow as tf
from tensorflow.keras import layers
from time import perf_counter
import os
from keras.callbacks import ModelCheckpoint
from keras.models import load_model
from tensorflow.keras.utils import plot_model
## Defining batch specfications
batch_size = 100
img_height = 250
img_width = 250
## loading training set
training_data = tf.keras.preprocessing.image_dataset_from_directory(
'D:/Dataset/Train',
seed=42,
image_size= (img_height, img_width),
batch_size=batch_size,
color_mode='rgb'
)
Found 231 files belonging to 2 classes.
## loading validation dataset
validation_data = tf.keras.preprocessing.image_dataset_from_directory(
'D:/Dataset/Val',
seed=42,
image_size= (img_height, img_width),
batch_size=batch_size,
color_mode='rgb'
)
Found 231 files belonging to 2 classes.
## loading testing dataset
testing_data = tf.keras.preprocessing.image_dataset_from_directory(
'D:/Dataset/test',
seed=42,
image_size= (img_height, img_width),
batch_size=batch_size,
color_mode='rgb'
)
Found 15 files belonging to 2 classes.
testing_data
<_BatchDataset element_spec=(TensorSpec(shape=(None, 250, 250, 3), dtype=tf.float32, name=None), TensorSpec(shape=(None,), dtype=tf.int32, name=None))>
class_names = training_data.class_names
class_names
['Training Non-Cancer', 'Training cancer']
## Configuring dataset for performance
AUTOTUNE = tf.data.experimental.AUTOTUNE
training_data = training_data.cache().prefetch(buffer_size=AUTOTUNE)
testing_data = testing_data.cache().prefetch(buffer_size=AUTOTUNE)
## Defining Cnn
model = tf.keras.models.Sequential([
layers.BatchNormalization(),
layers.Conv2D(32, 3, activation='relu'), # Conv2D(f_size, filter_size, activation) # relu, sigmoid, softmax
layers.MaxPooling2D(), # MaxPooling
layers.Conv2D(64, 3, activation='relu'),
layers.MaxPooling2D(),
layers.Conv2D(128, 3, activation='relu'),
layers.MaxPooling2D(),
layers.Conv2D(256, 3, activation='relu'),
layers.MaxPooling2D(),
layers.Flatten(),
layers.Dense(512, activation='relu'),
layers.Dense(len(class_names), activation= 'softmax')
])
model.compile(optimizer='adam',loss='sparse_categorical_crossentropy', metrics=['accuracy'])
model.build((None, 250, 250, 3))
model.summary()
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
batch_normalization (BatchN (None, 250, 250, 3) 12
ormalization)
conv2d (Conv2D) (None, 248, 248, 32) 896
max_pooling2d (MaxPooling2D (None, 124, 124, 32) 0
)
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
batch_normalization (BatchN (None, 250, 250, 3) 12
ormalization)
conv2d (Conv2D) (None, 248, 248, 32) 896
max_pooling2d (MaxPooling2D (None, 124, 124, 32) 0
)
conv2d_1 (Conv2D) (None, 122, 122, 64) 18496
max_pooling2d_1 (MaxPooling (None, 61, 61, 64) 0
2D)
conv2d_2 (Conv2D) (None, 59, 59, 128) 73856
max_pooling2d_2 (MaxPooling (None, 29, 29, 128) 0
2D)
conv2d_3 (Conv2D) (None, 27, 27, 256) 295168
max_pooling2d_3 (MaxPooling (None, 13, 13, 256) 0
2D)
flatten (Flatten) (None, 43264) 0
dense (Dense) (None, 512) 22151680
dense_1 (Dense) (None, 2) 1026
=================================================================
Total params: 22,541,134
Trainable params: 22,541,128
Non-trainable params: 6
_________________________________________________________________
## lets train our CNN
checkpoint = ModelCheckpoint("model_weights.h5", monitor='val_accuracy', verbose=1, save_best_only=True, mode='max')
callbacks_list = [checkpoint]
history = model.fit(training_data, validation_data=validation_data, epochs = 50, callbacks=callbacks_list)
Epoch 1/50 3/3 [==============================] - ETA: 0s - loss: 9.1437e-06 - accuracy: 1.0000 Epoch 1: val_accuracy improved from -inf to 0.54545, saving model to model_weights.h5 3/3 [==============================] - 30s 9s/step - loss: 9.1437e-06 - accuracy: 1.0000 - val_loss: 11.8699 - val_accuracy: 0.5455 Epoch 2/50 3/3 [==============================] - ETA: 0s - loss: 7.9455e-06 - accuracy: 1.0000 Epoch 2: val_accuracy did not improve from 0.54545 3/3 [==============================] - 20s 6s/step - loss: 7.9455e-06 - accuracy: 1.0000 - val_loss: 11.3084 - val_accuracy: 0.5455 Epoch 3/50 3/3 [==============================] - ETA: 0s - loss: 7.1059e-06 - accuracy: 1.0000 Epoch 3: val_accuracy did not improve from 0.54545 3/3 [==============================] - 20s 7s/step - loss: 7.1059e-06 - accuracy: 1.0000 - val_loss: 10.7544 - val_accuracy: 0.5455 Epoch 4/50 3/3 [==============================] - ETA: 0s - loss: 6.4954e-06 - accuracy: 1.0000 Epoch 4: val_accuracy did not improve from 0.54545 3/3 [==============================] - 16s 5s/step - loss: 6.4954e-06 - accuracy: 1.0000 - val_loss: 10.2117 - val_accuracy: 0.5455 Epoch 5/50 3/3 [==============================] - ETA: 0s - loss: 6.0331e-06 - accuracy: 1.0000 Epoch 5: val_accuracy did not improve from 0.54545 3/3 [==============================] - 16s 5s/step - loss: 6.0331e-06 - accuracy: 1.0000 - val_loss: 9.6825 - val_accuracy: 0.5455 Epoch 6/50 3/3 [==============================] - ETA: 0s - loss: 5.6755e-06 - accuracy: 1.0000 Epoch 6: val_accuracy did not improve from 0.54545 3/3 [==============================] - 16s 5s/step - loss: 5.6755e-06 - accuracy: 1.0000 - val_loss: 9.1676 - val_accuracy: 0.5455 Epoch 7/50 3/3 [==============================] - ETA: 0s - loss: 5.3917e-06 - accuracy: 1.0000 Epoch 7: val_accuracy did not improve from 0.54545 3/3 [==============================] - 17s 5s/step - loss: 5.3917e-06 - accuracy: 1.0000 - val_loss: 8.6670 - val_accuracy: 0.5455 Epoch 8/50 3/3 [==============================] - ETA: 0s - loss: 5.1636e-06 - accuracy: 1.0000 Epoch 8: val_accuracy did not improve from 0.54545 3/3 [==============================] - 16s 5s/step - loss: 5.1636e-06 - accuracy: 1.0000 - val_loss: 8.1808 - val_accuracy: 0.5455 Epoch 9/50 3/3 [==============================] - ETA: 0s - loss: 4.9747e-06 - accuracy: 1.0000 Epoch 9: val_accuracy did not improve from 0.54545 3/3 [==============================] - 17s 5s/step - loss: 4.9747e-06 - accuracy: 1.0000 - val_loss: 7.7089 - val_accuracy: 0.5455 Epoch 10/50 3/3 [==============================] - ETA: 0s - loss: 4.8173e-06 - accuracy: 1.0000 Epoch 10: val_accuracy did not improve from 0.54545 3/3 [==============================] - 17s 5s/step - loss: 4.8173e-06 - accuracy: 1.0000 - val_loss: 7.2512 - val_accuracy: 0.5455 Epoch 11/50 3/3 [==============================] - ETA: 0s - loss: 4.6873e-06 - accuracy: 1.0000 Epoch 11: val_accuracy did not improve from 0.54545 3/3 [==============================] - 16s 5s/step - loss: 4.6873e-06 - accuracy: 1.0000 - val_loss: 6.8084 - val_accuracy: 0.5455 Epoch 12/50 3/3 [==============================] - ETA: 0s - loss: 4.5712e-06 - accuracy: 1.0000 Epoch 12: val_accuracy improved from 0.54545 to 0.54978, saving model to model_weights.h5 3/3 [==============================] - 18s 6s/step - loss: 4.5712e-06 - accuracy: 1.0000 - val_loss: 6.3811 - val_accuracy: 0.5498 Epoch 13/50 3/3 [==============================] - ETA: 0s - loss: 4.4721e-06 - accuracy: 1.0000 Epoch 13: val_accuracy improved from 0.54978 to 0.55411, saving model to model_weights.h5 3/3 [==============================] - 18s 6s/step - loss: 4.4721e-06 - accuracy: 1.0000 - val_loss: 5.9703 - val_accuracy: 0.5541 Epoch 14/50 3/3 [==============================] - ETA: 0s - loss: 4.3807e-06 - accuracy: 1.0000 Epoch 14: val_accuracy improved from 0.55411 to 0.55844, saving model to model_weights.h5 3/3 [==============================] - 23s 8s/step - loss: 4.3807e-06 - accuracy: 1.0000 - val_loss: 5.5773 - val_accuracy: 0.5584 Epoch 15/50 3/3 [==============================] - ETA: 0s - loss: 4.2997e-06 - accuracy: 1.0000 Epoch 15: val_accuracy improved from 0.55844 to 0.58874, saving model to model_weights.h5 3/3 [==============================] - 26s 9s/step - loss: 4.2997e-06 - accuracy: 1.0000 - val_loss: 5.2029 - val_accuracy: 0.5887 Epoch 16/50 3/3 [==============================] - ETA: 0s - loss: 4.2280e-06 - accuracy: 1.0000 Epoch 16: val_accuracy improved from 0.58874 to 0.60606, saving model to model_weights.h5 3/3 [==============================] - 24s 9s/step - loss: 4.2280e-06 - accuracy: 1.0000 - val_loss: 4.8477 - val_accuracy: 0.6061 Epoch 17/50 3/3 [==============================] - ETA: 0s - loss: 4.1573e-06 - accuracy: 1.0000 Epoch 17: val_accuracy improved from 0.60606 to 0.61472, saving model to model_weights.h5 3/3 [==============================] - 20s 7s/step - loss: 4.1573e-06 - accuracy: 1.0000 - val_loss: 4.5111 - val_accuracy: 0.6147 Epoch 18/50 3/3 [==============================] - ETA: 0s - loss: 4.0918e-06 - accuracy: 1.0000 Epoch 18: val_accuracy improved from 0.61472 to 0.62771, saving model to model_weights.h5 3/3 [==============================] - 20s 7s/step - loss: 4.0918e-06 - accuracy: 1.0000 - val_loss: 4.1920 - val_accuracy: 0.6277 Epoch 19/50 3/3 [==============================] - ETA: 0s - loss: 4.0319e-06 - accuracy: 1.0000 Epoch 19: val_accuracy improved from 0.62771 to 0.63636, saving model to model_weights.h5 3/3 [==============================] - 19s 6s/step - loss: 4.0319e-06 - accuracy: 1.0000 - val_loss: 3.8899 - val_accuracy: 0.6364 Epoch 20/50 3/3 [==============================] - ETA: 0s - loss: 3.9725e-06 - accuracy: 1.0000 Epoch 20: val_accuracy improved from 0.63636 to 0.64502, saving model to model_weights.h5 3/3 [==============================] - 22s 8s/step - loss: 3.9725e-06 - accuracy: 1.0000 - val_loss: 3.6042 - val_accuracy: 0.6450 Epoch 21/50 3/3 [==============================] - ETA: 0s - loss: 3.9148e-06 - accuracy: 1.0000 Epoch 21: val_accuracy improved from 0.64502 to 0.66234, saving model to model_weights.h5 3/3 [==============================] - 21s 8s/step - loss: 3.9148e-06 - accuracy: 1.0000 - val_loss: 3.3344 - val_accuracy: 0.6623 Epoch 22/50 3/3 [==============================] - ETA: 0s - loss: 3.8580e-06 - accuracy: 1.0000 Epoch 22: val_accuracy improved from 0.66234 to 0.68398, saving model to model_weights.h5 3/3 [==============================] - 21s 7s/step - loss: 3.8580e-06 - accuracy: 1.0000 - val_loss: 3.0798 - val_accuracy: 0.6840 Epoch 23/50 3/3 [==============================] - ETA: 0s - loss: 3.8038e-06 - accuracy: 1.0000 Epoch 23: val_accuracy improved from 0.68398 to 0.68831, saving model to model_weights.h5 3/3 [==============================] - 19s 7s/step - loss: 3.8038e-06 - accuracy: 1.0000 - val_loss: 2.8400 - val_accuracy: 0.6883 Epoch 24/50 3/3 [==============================] - ETA: 0s - loss: 3.7491e-06 - accuracy: 1.0000 Epoch 24: val_accuracy improved from 0.68831 to 0.69697, saving model to model_weights.h5 3/3 [==============================] - 19s 6s/step - loss: 3.7491e-06 - accuracy: 1.0000 - val_loss: 2.6145 - val_accuracy: 0.6970 Epoch 25/50 3/3 [==============================] - ETA: 0s - loss: 3.6954e-06 - accuracy: 1.0000 Epoch 25: val_accuracy improved from 0.69697 to 0.72294, saving model to model_weights.h5 3/3 [==============================] - 19s 6s/step - loss: 3.6954e-06 - accuracy: 1.0000 - val_loss: 2.4039 - val_accuracy: 0.7229 Epoch 26/50 3/3 [==============================] - ETA: 0s - loss: 3.6371e-06 - accuracy: 1.0000 Epoch 26: val_accuracy improved from 0.72294 to 0.73593, saving model to model_weights.h5 3/3 [==============================] - 19s 7s/step - loss: 3.6371e-06 - accuracy: 1.0000 - val_loss: 2.2081 - val_accuracy: 0.7359 Epoch 27/50 3/3 [==============================] - ETA: 0s - loss: 3.5809e-06 - accuracy: 1.0000 Epoch 27: val_accuracy improved from 0.73593 to 0.75325, saving model to model_weights.h5 3/3 [==============================] - 21s 8s/step - loss: 3.5809e-06 - accuracy: 1.0000 - val_loss: 2.0268 - val_accuracy: 0.7532 Epoch 28/50 3/3 [==============================] - ETA: 0s - loss: 3.5184e-06 - accuracy: 1.0000 Epoch 28: val_accuracy did not improve from 0.75325 3/3 [==============================] - 17s 5s/step - loss: 3.5184e-06 - accuracy: 1.0000 - val_loss: 1.8581 - val_accuracy: 0.7532 Epoch 29/50 3/3 [==============================] - ETA: 0s - loss: 3.4550e-06 - accuracy: 1.0000 Epoch 29: val_accuracy improved from 0.75325 to 0.77056, saving model to model_weights.h5 3/3 [==============================] - 19s 7s/step - loss: 3.4550e-06 - accuracy: 1.0000 - val_loss: 1.7008 - val_accuracy: 0.7706 Epoch 30/50 3/3 [==============================] - ETA: 0s - loss: 3.3837e-06 - accuracy: 1.0000 Epoch 30: val_accuracy improved from 0.77056 to 0.78355, saving model to model_weights.h5 3/3 [==============================] - 20s 7s/step - loss: 3.3837e-06 - accuracy: 1.0000 - val_loss: 1.5542 - val_accuracy: 0.7835 Epoch 31/50 3/3 [==============================] - ETA: 0s - loss: 3.3053e-06 - accuracy: 1.0000 Epoch 31: val_accuracy improved from 0.78355 to 0.79654, saving model to model_weights.h5 3/3 [==============================] - 21s 7s/step - loss: 3.3053e-06 - accuracy: 1.0000 - val_loss: 1.4176 - val_accuracy: 0.7965 Epoch 32/50 3/3 [==============================] - ETA: 0s - loss: 3.2217e-06 - accuracy: 1.0000 Epoch 32: val_accuracy improved from 0.79654 to 0.80519, saving model to model_weights.h5 3/3 [==============================] - 21s 7s/step - loss: 3.2217e-06 - accuracy: 1.0000 - val_loss: 1.2901 - val_accuracy: 0.8052 Epoch 33/50 3/3 [==============================] - ETA: 0s - loss: 3.1216e-06 - accuracy: 1.0000 Epoch 33: val_accuracy improved from 0.80519 to 0.80952, saving model to model_weights.h5 3/3 [==============================] - 19s 6s/step - loss: 3.1216e-06 - accuracy: 1.0000 - val_loss: 1.1712 - val_accuracy: 0.8095 Epoch 34/50 3/3 [==============================] - ETA: 0s - loss: 3.0184e-06 - accuracy: 1.0000 Epoch 34: val_accuracy improved from 0.80952 to 0.81385, saving model to model_weights.h5 3/3 [==============================] - 19s 7s/step - loss: 3.0184e-06 - accuracy: 1.0000 - val_loss: 1.0598 - val_accuracy: 0.8139 Epoch 35/50 3/3 [==============================] - ETA: 0s - loss: 2.9054e-06 - accuracy: 1.0000 Epoch 35: val_accuracy improved from 0.81385 to 0.82684, saving model to model_weights.h5 3/3 [==============================] - 20s 7s/step - loss: 2.9054e-06 - accuracy: 1.0000 - val_loss: 0.9559 - val_accuracy: 0.8268 Epoch 36/50 3/3 [==============================] - ETA: 0s - loss: 2.7851e-06 - accuracy: 1.0000 Epoch 36: val_accuracy improved from 0.82684 to 0.83550, saving model to model_weights.h5 3/3 [==============================] - 22s 8s/step - loss: 2.7851e-06 - accuracy: 1.0000 - val_loss: 0.8591 - val_accuracy: 0.8355 Epoch 37/50 3/3 [==============================] - ETA: 0s - loss: 2.6654e-06 - accuracy: 1.0000 Epoch 37: val_accuracy improved from 0.83550 to 0.84416, saving model to model_weights.h5 3/3 [==============================] - 19s 7s/step - loss: 2.6654e-06 - accuracy: 1.0000 - val_loss: 0.7692 - val_accuracy: 0.8442 Epoch 38/50 3/3 [==============================] - ETA: 0s - loss: 2.5390e-06 - accuracy: 1.0000 Epoch 38: val_accuracy improved from 0.84416 to 0.85281, saving model to model_weights.h5 3/3 [==============================] - 19s 6s/step - loss: 2.5390e-06 - accuracy: 1.0000 - val_loss: 0.6860 - val_accuracy: 0.8528 Epoch 39/50 3/3 [==============================] - ETA: 0s - loss: 2.4177e-06 - accuracy: 1.0000 Epoch 39: val_accuracy improved from 0.85281 to 0.86580, saving model to model_weights.h5 3/3 [==============================] - 18s 6s/step - loss: 2.4177e-06 - accuracy: 1.0000 - val_loss: 0.6095 - val_accuracy: 0.8658 Epoch 40/50 3/3 [==============================] - ETA: 0s - loss: 2.2944e-06 - accuracy: 1.0000 Epoch 40: val_accuracy improved from 0.86580 to 0.87446, saving model to model_weights.h5 3/3 [==============================] - 18s 6s/step - loss: 2.2944e-06 - accuracy: 1.0000 - val_loss: 0.5396 - val_accuracy: 0.8745 Epoch 41/50 3/3 [==============================] - ETA: 0s - loss: 2.1757e-06 - accuracy: 1.0000 Epoch 41: val_accuracy improved from 0.87446 to 0.88312, saving model to model_weights.h5 3/3 [==============================] - 18s 6s/step - loss: 2.1757e-06 - accuracy: 1.0000 - val_loss: 0.4758 - val_accuracy: 0.8831 Epoch 42/50 3/3 [==============================] - ETA: 0s - loss: 2.0637e-06 - accuracy: 1.0000 Epoch 42: val_accuracy did not improve from 0.88312 3/3 [==============================] - 16s 5s/step - loss: 2.0637e-06 - accuracy: 1.0000 - val_loss: 0.4176 - val_accuracy: 0.8831 Epoch 43/50 3/3 [==============================] - ETA: 0s - loss: 1.9512e-06 - accuracy: 1.0000 Epoch 43: val_accuracy did not improve from 0.88312 3/3 [==============================] - 16s 5s/step - loss: 1.9512e-06 - accuracy: 1.0000 - val_loss: 0.3643 - val_accuracy: 0.8831 Epoch 44/50 3/3 [==============================] - ETA: 0s - loss: 1.8500e-06 - accuracy: 1.0000 Epoch 44: val_accuracy improved from 0.88312 to 0.90043, saving model to model_weights.h5 3/3 [==============================] - 17s 6s/step - loss: 1.8500e-06 - accuracy: 1.0000 - val_loss: 0.3157 - val_accuracy: 0.9004 Epoch 45/50 3/3 [==============================] - ETA: 0s - loss: 1.7499e-06 - accuracy: 1.0000 Epoch 45: val_accuracy improved from 0.90043 to 0.90909, saving model to model_weights.h5 3/3 [==============================] - 18s 6s/step - loss: 1.7499e-06 - accuracy: 1.0000 - val_loss: 0.2717 - val_accuracy: 0.9091 Epoch 46/50 3/3 [==============================] - ETA: 0s - loss: 1.6560e-06 - accuracy: 1.0000 Epoch 46: val_accuracy improved from 0.90909 to 0.92641, saving model to model_weights.h5 3/3 [==============================] - 18s 6s/step - loss: 1.6560e-06 - accuracy: 1.0000 - val_loss: 0.2321 - val_accuracy: 0.9264 Epoch 47/50 3/3 [==============================] - ETA: 0s - loss: 1.5662e-06 - accuracy: 1.0000 Epoch 47: val_accuracy improved from 0.92641 to 0.93939, saving model to model_weights.h5 3/3 [==============================] - 18s 6s/step - loss: 1.5662e-06 - accuracy: 1.0000 - val_loss: 0.1970 - val_accuracy: 0.9394 Epoch 48/50 3/3 [==============================] - ETA: 0s - loss: 1.4878e-06 - accuracy: 1.0000 Epoch 48: val_accuracy improved from 0.93939 to 0.95238, saving model to model_weights.h5 3/3 [==============================] - 17s 6s/step - loss: 1.4878e-06 - accuracy: 1.0000 - val_loss: 0.1663 - val_accuracy: 0.9524 Epoch 49/50 3/3 [==============================] - ETA: 0s - loss: 1.4119e-06 - accuracy: 1.0000 Epoch 49: val_accuracy improved from 0.95238 to 0.95671, saving model to model_weights.h5 3/3 [==============================] - 19s 6s/step - loss: 1.4119e-06 - accuracy: 1.0000 - val_loss: 0.1397 - val_accuracy: 0.9567 Epoch 50/50 3/3 [==============================] - ETA: 0s - loss: 1.3453e-06 - accuracy: 1.0000 Epoch 50: val_accuracy improved from 0.95671 to 0.96537, saving model to model_weights.h5 3/3 [==============================] - 18s 6s/step - loss: 1.3453e-06 - accuracy: 1.0000 - val_loss: 0.1171 - val_accuracy: 0.9654
###### serialize model structure to JSON
model_json = model.to_json()
with open("model.json", "w") as json_file:
json_file.write(model_json)
## stats on training data
plt.plot(history.history['loss'], label = 'training loss')
plt.plot(history.history['accuracy'], label = 'training accuracy')
plt.grid(True)
plt.legend()
<matplotlib.legend.Legend at 0x203c6226710>
## stats on training data
plt.plot(history.history['val_loss'], label = 'validation loss')
plt.plot(history.history['val_accuracy'], label = 'validation accuracy')
plt.grid(True)
plt.legend()
<matplotlib.legend.Legend at 0x20393822810>
## lets vizualize results on testing data
AccuracyVector = []
plt.figure(figsize=(30, 30))
for images, labels in testing_data.take(1):
predictions = model.predict(images)
predlabel = []
prdlbl = []
for mem in predictions:
predlabel.append(class_names[np.argmax(mem)])
prdlbl.append(np.argmax(mem))
AccuracyVector = np.array(prdlbl) == labels
for i in range(40):
ax = plt.subplot(10, 4, i + 1)
plt.imshow(images[i].numpy().astype("uint8"))
plt.title('Pred: '+ predlabel[i]+' actl:'+class_names[labels[i]] )
plt.axis('off')
plt.grid(True)
1/1 [==============================] - 0s 313ms/step
--------------------------------------------------------------------------- InvalidArgumentError Traceback (most recent call last) Cell In[45], line 16 14 for i in range(40): 15 ax = plt.subplot(10, 4, i + 1) ---> 16 plt.imshow(images[i].numpy().astype("uint8")) 17 plt.title('Pred: '+ predlabel[i]+' actl:'+class_names[labels[i]] ) 18 plt.axis('off') File ~\AppData\Roaming\Python\Python311\site-packages\tensorflow\python\util\traceback_utils.py:153, in filter_traceback.<locals>.error_handler(*args, **kwargs) 151 except Exception as e: 152 filtered_tb = _process_traceback_frames(e.__traceback__) --> 153 raise e.with_traceback(filtered_tb) from None 154 finally: 155 del filtered_tb File ~\AppData\Roaming\Python\Python311\site-packages\tensorflow\python\framework\ops.py:7262, in raise_from_not_ok_status(e, name) 7260 def raise_from_not_ok_status(e, name): 7261 e.message += (" name: " + name if name is not None else "") -> 7262 raise core._status_to_exception(e) from None InvalidArgumentError: {{function_node __wrapped__StridedSlice_device_/job:localhost/replica:0/task:0/device:CPU:0}} slice index 15 of dimension 0 out of bounds. [Op:StridedSlice] name: strided_slice/
from keras.models import model_from_json
import numpy as np
class cancerDetectionModel(object):
class_nums = ['Training cancer', "Training Non-Cancer"]
def __init__(self, model_json_file, model_weights_file):
# load model from JSON file
with open(model_json_file, "r") as json_file:
loaded_model_json = json_file.read()
self.loaded_model = model_from_json(loaded_model_json)
# load weights into the new model
self.loaded_model.load_weights(model_weights_file)
self.loaded_model.make_predict_function()
def predict_cancer(self, img):
self.preds = self.loaded_model.predict(img)
return cancerDetectionModel.class_nums[np.argmax(self.preds)], self.preds
full_train_df = pd.read_csv("D:/Dataset/converted_keras/labels.txt")
full_train_df.head()
| 0 cancer | |
|---|---|
| 0 | 1 Non-Cancer |
print("Train Size: {}".format(len(os.listdir('D:\Dataset\Train\Training cancer'))))
print("Test Size: {}".format(len(os.listdir('D:\Dataset\Train\Training Non-Cancer'))))
Train Size: 126 Test Size: 105
import os
from PIL import Image
from matplotlib import pyplot as plt
root1 = 'D:\Dataset\Train\Training cancer'
fnames = os.listdir(root1)
len(fnames)
105
root2 = 'D:\Dataset\Train\Training Non-Cancer'
fnames = os.listdir(root2)
len(fnames)
105
fig , axs = plt.subplots(nrows=2,ncols=5,figsize=(10,10))
axs = axs.flatten()
for i in range(10):
filepath = os.path.join(root1,root2,fnames[i])
img = Image.open(filepath)
axs[i].imshow(img)
axs[i].axis('off')
axs[i].set_title(fnames[i])
plt.show()
##### For CSV FORMATTTT
# from PIL import Image
# import numpy as np
# import sys
# import os
# import csv
# def createFileList(myDir, format='jpg'):
# fileList = []
# print(myDir)
# for root, dirs, files in os.walk(myDir, topdown = False):
# for name in files:
# if name.endswith(format):
# fullName = os.path.join(root, name)
# fileList.append(fullName)
# return fileList
# myFileList = createFileList('D:\Dataset\Train\Training cancer')
# for file in myFileList:
# print(file)
# img_file = Image.open(file)
# width, height = img_file.size
# format = img_file.format
# mode = img_file.mode
# img_grey = img_file.convert('L')
# value = np.asarray(img_grey.getdata(), dtype=np.int).reshape((img_grey.size[1], img_grey.size[0]))
# value = value.flatten()
# print(value)
# with open("image_to_csv.csv", 'a') as f:
# writer = csv.writer(f)
# writer.writerow(value)
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
import os
path = os.listdir('D:/Dataset/Train')
classes = {'Training cancer': 0, 'Training Non-Cancer': 1}
import cv2
X = []
Y = []
for cls in classes:
pth = 'D:/Dataset/Train/'+cls
for j in os.listdir(pth):
img = cv2.imread(pth+'/'+j,0)
img = cv2.resize(img, (1028,1028))
X.append(img)
Y.append(classes[cls])
np.unique(Y)
array([0, 1])
X = np.array(X)
Y = np.array(Y)
pd.Series(Y).value_counts()
1 290 0 251 dtype: int64
X.shape
(541, 1028, 1028)
#Visualize Data
plt.imshow(X[0], cmap='gray')
<matplotlib.image.AxesImage at 0x17d5d7e8e10>
#Prepare Data
X_updated = X.reshape(len(X), -1)
X_updated.shape
(541, 1056784)
#split Data
xtrain, xtest, ytrain, ytest = train_test_split(X_updated, Y, random_state=10,test_size=.20)
xtrain.shape, xtest.shape
((432, 1056784), (109, 1056784))
#Feature Scaling
print(xtrain.max(), xtrain.min())
print(xtest.max(), xtest.min())
xtrain = xtrain/255
xtest = xtest/255
print(xtrain.max(), xtrain.min())
print(xtest.max(), xtest.min())
255 5 255 7 1.0 0.0196078431372549 1.0 0.027450980392156862
#Fearute selection: PCA
from sklearn.decomposition import PCA
print(xtrain.shape, xtest.shape)
pca = PCA(.98)
# pca_train = pca.fit_transform(xtrain)
# pca_test = pca.transform(xtest)
pca_train = xtrain
pca_test = xtest
(432, 1056784) (109, 1056784)
# print(pca_train.shape, pca_test.shape)
# print(pca.n_components_)
# print(pca.n_features_)
###TRAIN MODEL
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
import warnings
warnings.filterwarnings('ignore')
lg = LogisticRegression(C=0.1) #Penality Parameter
lg.fit(pca_train, ytrain)
LogisticRegression(C=0.1)In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
LogisticRegression(C=0.1)
sv = SVC()
sv.fit(pca_train, ytrain)
SVC()In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
SVC()
#Evaluation
print("Training Score :", lg.score(pca_train, ytrain))
print("Testing Score:", lg.score(pca_test, ytest))
Training Score : 1.0 Testing Score: 0.9724770642201835
print("Training Score :", sv.score(pca_train, ytrain))
print("Testing Score:", sv.score(pca_test, ytest))
Training Score : 1.0 Testing Score: 0.9724770642201835
#Prediction
pred = sv.predict(pca_test)
np.where(ytest!=pred)
(array([ 22, 59, 102], dtype=int64),)
pred[36]
1
ytest[36]
1
#TEST MODEL
dec = {0:'cancer', 1:'Non-Cancer'}
plt.figure(figsize=(12,8))
p = os.listdir('D:/Dataset/test/')
c=1
for i in os.listdir('D:/Dataset/test/cancer/')[:8]:
plt.subplot(3,3,c)
img = cv2.imread('D:/Dataset/test/cancer/'+i,0)
img1 = cv2.resize(img, (1028,1028))
img1 = img1.reshape(1,-1)/255
p = sv.predict(img1)
plt.title(dec[p[0]])
plt.imshow(img, cmap='gray')
plt.axis('off')
c+=1
plt.figure(figsize=(12,8))
p = os.listdir('D:/Dataset/test/')
c=1
for i in os.listdir('D:/Dataset/test/Non-Cancer/')[:6]:
plt.subplot(3,3,c)
img = cv2.imread('D:/Dataset/test/Non-Cancer/'+i,0)
img1 = cv2.resize(img, (1028,1028))
img1 = img1.reshape(1,-1)/255
p = sv.predict(img1)
plt.title(dec[p[0]])
plt.imshow(img, cmap='gray')
plt.axis('off')
c+=1
BASE_PATH = 'D:/Dataset/Train/'
unique_classes = []
for path in os.listdir(BASE_PATH):
unique_classes.append(path)
print(unique_classes)
['Training cancer', 'Training Non-Cancer']
class_index = [unique_classes[1], unique_classes[0]]
for c in class_index:
print(c, "-", class_index.index(c))
Training Non-Cancer - 0 Training cancer - 1
images = []
masks = []
labels = []
for folder in os.listdir(BASE_PATH):
class_path = os.path.join(BASE_PATH, folder)
for img in os.listdir(class_path):
if "_mask" not in img:
img_path = os.path.join(class_path, img)
msk_path = img_path.replace(".png", "_mask.png")
# check if mask exist
if os.path.exists(msk_path):
images.append(img_path)
masks.append(msk_path)
labels.append(folder)
print(len(images))
541
images[0]
'D:/Dataset/Train/Training cancer\\1056.jpg'
input_images_size = 256
channel = 1
import cv2
import scipy
import scipy.ndimage
def load_image(img_path):
""" Load single image as Grayscale
"""
# load image as grayscale
img = cv2.imread(img_path, 0)
return img
def padding(img, msk):
""" Pad images to make them square
"""
size = np.max(img.shape)
offset_x = (size-img.shape[0])//2
offset_y = (size-img.shape[1])//2
blank_image = np.zeros((size, size))
blank_mask = np.zeros((size, size))
blank_image[offset_x:offset_x+img.shape[0],
offset_y:offset_y+img.shape[1]] = img
blank_mask[offset_x:offset_x+img.shape[0],
offset_y:offset_y+img.shape[1]] = msk
return blank_image, blank_mask
def resize_mask(mask):
"""Resize mask, its different because mask pixel value can change because of resize
"""
new_size = np.array([input_images_size, input_images_size]) / mask.shape
mask = scipy.ndimage.interpolation.zoom(mask, new_size)
return mask
def resize(img):
"""Resize image
"""
img = cv2.resize(img, (input_images_size, input_images_size))
return img
def preprocess(img):
"""Image preprocessing
Normalize image
"""
img = img/255.0
return img
def inverse_preprocess(img):
"""Inverse of preprocessing
"""
img = img*255
return img
def load_data(img_path, msk_path, label):
"""Load image, mask and repalce mask value with class index
0 = normal
1 = benign
2 = malignant
"""
img = load_image(img_path)
msk = load_image(msk_path)
img, msk = padding(img, msk)
label_indx = class_index.index(label)
msk[msk == 255] = 1
msk = msk.astype("uint8")
img = resize(img)
msk = resize_mask(msk)
new_mask = np.zeros((input_images_size, input_images_size, 2))
if label_indx != 0:
new_mask[:, :, label_indx-1] = msk
# print(np.unique(msk), label, label_indx)
return img, new_mask
def load_batch(images, masks, labels):
"""Load Batch of data
"""
batch_x = []
batch_y = []
for i, m, l in zip(images, masks, labels):
img, msk = load_data(i, m, l)
img = preprocess(img)
batch_x.append(img)
batch_y.append(msk)
return np.array(batch_x), np.array(batch_y)
import matplotlib.pyplot as plt
for i in [0, 500, 600]:
indx = i
img, msk = load_data(images[indx], masks[indx], labels[indx])
print(np.min(img), np.max(img), img.shape)
plt.figure(figsize=(20, 20))
plt.subplot(1, 3, 1)
plt.axis("off")
plt.imshow(img)
plt.subplot(1, 3, 2)
plt.axis("off")
plt.imshow(msk[:, :, 0])
plt.subplot(1, 3, 3)
plt.axis("off")
plt.imshow(msk[:, :, 1])
plt.show()
16.55059814453125 217.49298095703125 (256, 256)
21.66058349609375 251.78179931640625 (256, 256)
--------------------------------------------------------------------------- IndexError Traceback (most recent call last) Cell In[52], line 4 2 for i in [0, 500, 600]: 3 indx = i ----> 4 img, msk = load_data(images[indx], masks[indx], labels[indx]) 5 print(np.min(img), np.max(img), img.shape) 6 plt.figure(figsize=(20, 20)) IndexError: list index out of range
images = np.array(images)
masks = np.array(masks)
labels = np.array(labels)
!pip install segmentation_models
Requirement already satisfied: segmentation_models in c:\python311\lib\site-packages (1.0.1) Requirement already satisfied: keras-applications<=1.0.8,>=1.0.7 in c:\python311\lib\site-packages (from segmentation_models) (1.0.8) Requirement already satisfied: image-classifiers==1.0.0 in c:\python311\lib\site-packages (from segmentation_models) (1.0.0) Requirement already satisfied: efficientnet==1.0.0 in c:\python311\lib\site-packages (from segmentation_models) (1.0.0) Requirement already satisfied: scikit-image in c:\python311\lib\site-packages (from efficientnet==1.0.0->segmentation_models) (0.20.0) Requirement already satisfied: numpy>=1.9.1 in c:\python311\lib\site-packages (from keras-applications<=1.0.8,>=1.0.7->segmentation_models) (1.23.5) Requirement already satisfied: h5py in c:\users\hp\appdata\roaming\python\python311\site-packages (from keras-applications<=1.0.8,>=1.0.7->segmentation_models) (3.8.0) Requirement already satisfied: scipy>=1.8 in c:\python311\lib\site-packages (from scikit-image->efficientnet==1.0.0->segmentation_models) (1.10.1) Requirement already satisfied: networkx>=2.8 in c:\python311\lib\site-packages (from scikit-image->efficientnet==1.0.0->segmentation_models) (3.1) Requirement already satisfied: pillow>=9.0.1 in c:\python311\lib\site-packages (from scikit-image->efficientnet==1.0.0->segmentation_models) (9.4.0) Requirement already satisfied: imageio>=2.4.1 in c:\python311\lib\site-packages (from scikit-image->efficientnet==1.0.0->segmentation_models) (2.28.1) Requirement already satisfied: tifffile>=2019.7.26 in c:\python311\lib\site-packages (from scikit-image->efficientnet==1.0.0->segmentation_models) (2023.4.12) Requirement already satisfied: PyWavelets>=1.1.1 in c:\python311\lib\site-packages (from scikit-image->efficientnet==1.0.0->segmentation_models) (1.4.1) Requirement already satisfied: packaging>=20.0 in c:\python311\lib\site-packages (from scikit-image->efficientnet==1.0.0->segmentation_models) (23.0) Requirement already satisfied: lazy_loader>=0.1 in c:\python311\lib\site-packages (from scikit-image->efficientnet==1.0.0->segmentation_models) (0.2)
pip install upgrade keras
Note: you may need to restart the kernel to use updated packages.
ERROR: Could not find a version that satisfies the requirement upgrade (from versions: none) ERROR: No matching distribution found for upgrade
import os
os.environ["SM_FRAMEWORK"] = "tf.keras"
from tensorflow import keras
import segmentation_models as sm
import tensorflow as tf
sm.framework()
BACKBONE = 'resnet34'
LR = 0.00001
model = sm.Unet(BACKBONE, classes=2, activation="sigmoid",input_shape=(input_images_size,input_images_size, channel),encoder_weights=None)
optim = tf.keras.optimizers.Adam(LR)
dice_loss = sm.losses.DiceLoss()
focal_loss = sm.losses.BinaryFocalLoss()
total_loss = dice_loss + (1 * focal_loss)
metrics = [sm.metrics.IOUScore(threshold=0.5),
sm.metrics.FScore(threshold=0.5)]
model.compile(optim, total_loss, metrics)
Segmentation Models: using `tf.keras` framework.
batch_size = 4
history = {"epoch": []}
for e in range(100):
print("epoch:",e, end=" > ")
indexes = list(range(len(images)))
temp_history = {"loss": [],
"IOU": [],
"F-Score": []}
for b in range(0, len(images), batch_size):
bs = b
be = bs+batch_size
batch_index = indexes[bs:be]
batch_x, batch_y = load_batch(images[batch_index], masks[batch_index], labels[batch_index])
batch_x = np.expand_dims(batch_x, axis=-1)
batch_y = np.expand_dims(batch_y, axis=-1)
batch_y = batch_y.astype("float32")
loss = model.train_on_batch(batch_x, batch_y)
temp_history["loss"].append(loss[0])
temp_history["IOU"].append(loss[1])
temp_history["F-Score"].append(loss[2])
print("loss", np.round(np.mean(temp_history["loss"]), 4),"IOU", np.round(np.mean(temp_history["IOU"]), 4),"F-Score", np.round(np.mean(temp_history["F-Score"]), 4))
history["epoch"].append(temp_history)
model.save_weights("cancer")
epoch: 0 > loss 0.814 IOU 0.1356 F-Score 0.1698 epoch: 1 > loss -1.2409 IOU 0.2723 F-Score 0.2485 epoch: 2 > loss -4.4345 IOU 0.3412 F-Score 0.2755 epoch: 3 > loss -8.0726 IOU 0.2924 F-Score 0.258 epoch: 4 > loss -11.4899 IOU 0.2444 F-Score 0.2374 epoch: 5 >
--------------------------------------------------------------------------- KeyboardInterrupt Traceback (most recent call last) Cell In[66], line 17 15 batch_y = np.expand_dims(batch_y, axis=-1) 16 batch_y = batch_y.astype("float32") ---> 17 loss = model.train_on_batch(batch_x, batch_y) 18 temp_history["loss"].append(loss[0]) 19 temp_history["IOU"].append(loss[1]) File c:\Python311\Lib\site-packages\keras\engine\training.py:2510, in Model.train_on_batch(self, x, y, sample_weight, class_weight, reset_metrics, return_dict) 2506 iterator = data_adapter.single_batch_iterator( 2507 self.distribute_strategy, x, y, sample_weight, class_weight 2508 ) 2509 self.train_function = self.make_train_function() -> 2510 logs = self.train_function(iterator) 2512 logs = tf_utils.sync_to_numpy_or_python_type(logs) 2513 if return_dict: File ~\AppData\Roaming\Python\Python311\site-packages\tensorflow\python\util\traceback_utils.py:150, in filter_traceback.<locals>.error_handler(*args, **kwargs) 148 filtered_tb = None 149 try: --> 150 return fn(*args, **kwargs) 151 except Exception as e: 152 filtered_tb = _process_traceback_frames(e.__traceback__) File ~\AppData\Roaming\Python\Python311\site-packages\tensorflow\python\eager\polymorphic_function\polymorphic_function.py:894, in Function.__call__(self, *args, **kwds) 891 compiler = "xla" if self._jit_compile else "nonXla" 893 with OptionalXlaContext(self._jit_compile): --> 894 result = self._call(*args, **kwds) 896 new_tracing_count = self.experimental_get_tracing_count() 897 without_tracing = (tracing_count == new_tracing_count) File ~\AppData\Roaming\Python\Python311\site-packages\tensorflow\python\eager\polymorphic_function\polymorphic_function.py:926, in Function._call(self, *args, **kwds) 923 self._lock.release() 924 # In this case we have created variables on the first call, so we run the 925 # defunned version which is guaranteed to never create variables. --> 926 return self._no_variable_creation_fn(*args, **kwds) # pylint: disable=not-callable 927 elif self._variable_creation_fn is not None: 928 # Release the lock early so that multiple threads can perform the call 929 # in parallel. 930 self._lock.release() File ~\AppData\Roaming\Python\Python311\site-packages\tensorflow\python\eager\polymorphic_function\tracing_compiler.py:143, in TracingCompiler.__call__(self, *args, **kwargs) 140 with self._lock: 141 (concrete_function, 142 filtered_flat_args) = self._maybe_define_function(args, kwargs) --> 143 return concrete_function._call_flat( 144 filtered_flat_args, captured_inputs=concrete_function.captured_inputs) File ~\AppData\Roaming\Python\Python311\site-packages\tensorflow\python\eager\polymorphic_function\monomorphic_function.py:1757, in ConcreteFunction._call_flat(self, args, captured_inputs, cancellation_manager) 1753 possible_gradient_type = gradients_util.PossibleTapeGradientTypes(args) 1754 if (possible_gradient_type == gradients_util.POSSIBLE_GRADIENT_TYPES_NONE 1755 and executing_eagerly): 1756 # No tape is watching; skip to running the function. -> 1757 return self._build_call_outputs(self._inference_function.call( 1758 ctx, args, cancellation_manager=cancellation_manager)) 1759 forward_backward = self._select_forward_and_backward_functions( 1760 args, 1761 possible_gradient_type, 1762 executing_eagerly) 1763 forward_function, args_with_tangents = forward_backward.forward() File ~\AppData\Roaming\Python\Python311\site-packages\tensorflow\python\eager\polymorphic_function\monomorphic_function.py:381, in _EagerDefinedFunction.call(self, ctx, args, cancellation_manager) 379 with _InterpolateFunctionError(self): 380 if cancellation_manager is None: --> 381 outputs = execute.execute( 382 str(self.signature.name), 383 num_outputs=self._num_outputs, 384 inputs=args, 385 attrs=attrs, 386 ctx=ctx) 387 else: 388 outputs = execute.execute_with_cancellation( 389 str(self.signature.name), 390 num_outputs=self._num_outputs, (...) 393 ctx=ctx, 394 cancellation_manager=cancellation_manager) File ~\AppData\Roaming\Python\Python311\site-packages\tensorflow\python\eager\execute.py:52, in quick_execute(op_name, num_outputs, inputs, attrs, ctx, name) 50 try: 51 ctx.ensure_initialized() ---> 52 tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name, 53 inputs, attrs, num_outputs) 54 except core._NotOkStatusException as e: 55 if name is not None: KeyboardInterrupt:
import matplotlib.pyplot as plt
for i in [0, 500, 600]:
indx = i
img, msk = load_data(images[indx], masks[indx], labels[indx])
print(np.min(img), np.max(img), img.shape)
print(img.shape)
img2 = preprocess(img)
pred = model.predict(np.array([img2]))
pred = pred[0]
plt.figure(figsize=(20, 20))
plt.subplot(1, 3, 1)
plt.axis("off")
plt.imshow(img)
plt.subplot(1, 3, 2)
plt.axis("off")
plt.imshow(pred[:, :, 0])
plt.subplot(1, 3, 3)
plt.axis("off")
plt.imshow(pred[:, :, 1])
plt.show()
16.55059814453125 217.49298095703125 (256, 256) (256, 256) 1/1 [==============================] - 1s 1s/step
21.66058349609375 251.78179931640625 (256, 256) (256, 256) 1/1 [==============================] - 0s 160ms/step
--------------------------------------------------------------------------- IndexError Traceback (most recent call last) Cell In[67], line 4 2 for i in [0, 500, 600]: 3 indx = i ----> 4 img, msk = load_data(images[indx], masks[indx], labels[indx]) 5 print(np.min(img), np.max(img), img.shape) 6 print(img.shape) IndexError: index 600 is out of bounds for axis 0 with size 541
from keras.models import load_model # TensorFlow is required for Keras to work
from PIL import Image, ImageOps # Install pillow instead of PIL
import numpy as np
# Disable scientific notation for clarity
np.set_printoptions(suppress=True)
# Load the model
model = load_model("D:/Dataset/converted_keras/keras_model.h5", compile=False)
# Load the labels
class_names = open("D:/Dataset/converted_keras/labels.txt", "r").readlines()
# Create the array of the right shape to feed into the keras model
# The 'length' or number of images you can put into the array is
# determined by the first position in the shape tuple, in this case 1
data = np.ndarray(shape=(1, 224, 224, 3), dtype=np.float32)
# Replace this with the path to your image ... PATH...PATH..PATH
image = Image.open("D:/Dataset/test/Non-Cancer/6.png").convert("RGB")
# resizing the image to be at least 224x224 and then cropping from the center
size = (224, 224)
image = ImageOps.fit(image, size, Image.Resampling.LANCZOS)
# turn the image into a numpy array
image_array = np.asarray(image)
# Normalize the image
normalized_image_array = (image_array.astype(np.float32) / 127.5) - 1
# Load the image into the array
data[0] = normalized_image_array
# Predicts the model
prediction = model.predict(data)
index = np.argmax(prediction)
class_name = class_names[index]
confidence_score = prediction[0][index]
# Print prediction and confidence score
print("Class:", class_name[2:], end="")
print("Confidence Score:", confidence_score)
1/1 [==============================] - 1s 1s/step Class: cancer Confidence Score: 0.9686299